Environmental Decision-Making - New Ecoinformatics Tools in Environmental Science: Applications and Decision-making - page 238

airborne and fl floating laboratories, and stationary systems. In spite of the possibility

of de nite realization of this structure and the applicable computer technique, there

exists, nevertheless, one general problem whose solution determines the detection
ef

ciency. This task involves the organization of in-line data processing at all
detection system levels under real-time conditions, ensuring an uninterrupted
matching of the operation of all the cybernetic devices of the system and also taking
into account the limitation on their functional characteristics.
Statistical analysis of the information at each level of the monitoring system is
completed by making the following two subsequent decisions:

(1) storage of features of the landscape elements for the accumulation of addi-
tional information about them,
(2) completion on the presence of an anomaly and transmission of its character-
istics to the next data processing level.

rst stage
selection and storage of parameters of the elements of landscapes suspected of
being anomalous are dependent on the character of the algorithm used for singling
out the two-dimensional signal against the background noise. They are also asso-
ciated with the gradual

These two stages determine the ef cacy of the detection system. At the

filling of the main storage. The second stage determines the
magnitude of the detection probability, depending on the criteria of registration of
the landscape elements selected at the first stage for further analysis. As a conse-
quence, the problem of matching the flow of information between the processing

stages under monitoring conditions and at minimal losses arises. The solution to
this problem involves two procedures of parallel component processing, forming
the intermediate information delay either as a constant computer memory buffer
size or for a

fl

fixed storage time.

3.5.2 Multi-channel Statistical Analyzer

Real monitoring systems have series of informational channels the ef ciency of

whom depends on the solved task. Usually, the principal scheme of monitoring
system has the form of some transformation F. Its structure and functions are
de ned by the items realizing separate operations on the analysis and transforma-

tion of measurement data. This is given schematically in Fig. 3.9 .
The general scheme of sensing the environmental elements assumes that the

ux
of data {T b } from the j-th measuring device can be analyzed both separately and
together with the data from other channels. In fact, there is a digital processor of the
type of a personal computer with a speed equal to V operations per second. The task
is set to identify an unknown vector parameter T b ={T b1 ,

fl

…

, T bn } characterizing
the state of the controlled object (e.g., in the problems of microwave monitoring`
these are brightness temperatures). To solve this problem, an algorithm of the
parallel-in-time analysis of the vector T b components is proposed, which makes it